Principal component and Tucker3 analyses of high performance liquid chromatography with diode-array detection fingerprints of crude extracts of Erythrina speciosa Andrews leaves Original Research Article

Mixtures of ethanol, dichloromethane, hexane and acetone obtained according to a statistical design have been used to extract substances from Erythrina speciosa Andrew leaves for chromatographic fingerprinting. The plant extracts from each mixture were analyzed by HPLC-DAD providing UV–vis spectra for each chromatographic peak. These chromatograms and spectra for the design mixtures were then treated with principal component (PCA), Tucker3 and PARAFAC analyses. PCA indicated the existence of five different chromatographic fingerprints for the leave extracts depending on the solvent mixture composition. Different chromatographic peak areas were strongly correlated with the mixture proportions of acetone, dichloromethane and ethanol. Tucker3 and PARAFAC analyses were very useful for identifying simultaneous correlations between chromatographic peak areas, spectral band absorbances and solvent proportions. The acetone proportion was highly correlated with the area of the 3.69 min retention time peak and the spectral absorbances between 250 and 260 nm, consistent with the presence of natural polyphenols. The dichloromethane mixture proportion was strongly correlated with the 12.19 min chromatographic peak area and a single spectral absorbance at 201 nm. This spectral absorption is characteristic of the electronic structures of terpenes and alkaloids.